172  So-Called  Tasteless  Iron  Compounds.  \Km^^T' 
to  form  double  salts,  the  latter  property  is  exerted  before  entire  double 
decomposition  is  effected  ;  hence,  in  order  to  insure  this  completely, 
sufficient  monad  citrate  must  be  present  to  form  double  salt  with  the 
ferric  citrate  to  be  produced.  This  is  very  clearly  seen  from  the  fol- 
lowing equation  : 
Fe2Cl6+4(K3C6H507)=6(KCl)+2(FeC6H5OrK3C6H607). 
On  evaporating  the  solution  containing  the  mixture  represented  by 
the  second  member  of  this  equation,  a  greenish-white  granular  mass  is 
obtained.  Since  potassium  chloride  is  not  sufficiently  soluble  in  alco- 
hol to  effect  a  separation,  the  writer  proposed  to  eliminate  it  by  means 
of  dialysis.  A  capacious  parchment  filter,  containing  the  concentrated 
solution,  was  suspended  in  pure  water,  and  after  a  few  moments  a 
strong  exosmotic  current  of  heavy  liquid  was  seen  descending  from 
the  apex  of  the  cone.  After  about  six  hours,  the  powerful  endosmotic 
current  had  nearly  filled  the  dialyser  which  at  first  was  scarcely  one- 
third  full.  The  diffusate  had  assumed  a  yellowish  tint,  showing  that  a 
trace  of  iron  had  also  passed  ;  but  on  concentration  the  potassium 
chloride  began  to  crystallize  abundantly,  and  on  evaporation  to  dryness 
a  crystalline  residue  of  potassium  chloride  was  obtained,  corresponding 
to  over  §  of  the  theoretical  yield.  Having  found  the  parchment  filter 
too  inconvenient,  the  writer  again  concentrated  the  liquid  in  the  dia- 
lyser, placed  it  into  a  capacious  wide-mouthed  bottle,  then  tied  over 
the  mouth  with  pig's  bladder  previously  washed  in  very  dilute  ammo- 
nia, and  inverted  the  bottle  into  a  vessel  containing  pure  water.  The 
dialysis  now  proceeded  more  satisfactorily,  since  the  endosmotic  cur- 
rent was  nearly  overcome  by  the  tension  of  the  atmosphere  within  the 
bottle.  A  slight  increase  of  the  volume  of  the  diffusum  was,  how- 
ever, noticed,  and  the  consequent  compression  of  the  enclosed  air  dis- 
tended the  membrane  considerably. 
The  dialysis  was  discontinued  after  about  48  hours,  having  in  the 
mean  time  changed  the  diffusate  frequently.  The  colloidal  remnant  of  the 
diffusum  was  then  tested  for  chlorine  with  argentic  nitrate,  after  strong 
acidulation  with  nitric  acid,  when  only  an  opalescence  appeared,  show- 
ing that  complete  removal  of  the  potassium  chloride  had  taken  place, 
whilst  the  green  color  of  the  residual  liquid  remained  unchanged. 
The  iodide  was  then  subjected  to  dialysis.  As  the  writer  had  already 
separated  the  potassium  iodide  by  means  of  alcohol,  a  trial  with  dialy- 
sis seemed  really  superfluous  ;  but,  as  it  had  been  intimated  that  free 
